openai

Compare original and translation side by side

🇺🇸

Original

English

🇨🇳

Translation

Chinese

OpenAI API

Use the OpenAI API via direct

curl

calls to access GPT models, DALL-E image generation, Whisper transcription, embeddings, and text-to-speech.

Official docs:

https://platform.openai.com/docs/api-reference

通过直接调用

curl

来使用OpenAI API，访问GPT模型、DALL-E图像生成、Whisper转写、文本嵌入以及文本转语音功能。

官方文档：

https://platform.openai.com/docs/api-reference

When to Use

使用场景

Use this skill when you need to:

Chat completions with GPT-4o, GPT-4, or GPT-3.5 models
Image generation with DALL-E 3
Audio transcription with Whisper
Text-to-speech audio generation
Text embeddings for semantic search and RAG
Vision tasks (analyze images with GPT-4o)

在以下场景中可使用该技能：

使用GPT-4o、GPT-4或GPT-3.5模型进行对话补全
使用DALL-E 3进行图像生成
使用Whisper进行音频转写
生成文本转语音音频
用于语义搜索和RAG的文本嵌入
视觉任务（使用GPT-4o分析图像）

Prerequisites

前置条件

Sign up at OpenAI Platform and create an account
Go to API Keys and generate a new secret key
Add billing information and set usage limits

bash

export OPENAI_API_KEY="sk-..."

在OpenAI平台注册并创建账号
前往API密钥页面生成新的密钥
添加账单信息并设置使用限额

bash

export OPENAI_API_KEY="sk-..."

Pricing (as of 2025)

定价（截至2025年）

Model	Input (per 1M tokens)	Output (per 1M tokens)
GPT-4o	$2.50	$10.00
GPT-4o-mini	$0.15	$0.60
GPT-4 Turbo	$10.00	$30.00
text-embedding-3-small	$0.02	-
text-embedding-3-large	$0.13	-

模型	输入（每100万tokens）	输出（每100万tokens）
GPT-4o	$2.50	$10.00
GPT-4o-mini	$0.15	$0.60
GPT-4 Turbo	$10.00	$30.00
text-embedding-3-small	$0.02	-
text-embedding-3-large	$0.13	-

Rate Limits

速率限制

Rate limits vary by tier (based on usage history). Check your limits at Platform Settings.

Important: When using
$VAR
in a command that pipes to another command, wrap the command containing
$VAR
in
bash -c '...'
. Due to a Claude Code bug, environment variables are silently cleared when pipes are used directly.
bash
bash -c 'curl -s "https://api.example.com" -H "Authorization: Bearer $API_KEY"' | jq .

速率限制因等级而异（基于使用历史）。可在平台设置页面查看你的限额。

重要提示： 当在包含管道的命令中使用
$VAR
时，请将包含
$VAR
的命令用
bash -c '...'
包裹。由于Claude Code的bug，直接使用管道时环境变量会被自动清除。
bash
bash -c 'curl -s "https://api.example.com" -H "Authorization: Bearer $API_KEY"' | jq .

How to Use

使用方法

All examples below assume you have

OPENAI_API_KEY

set.

Base URL:

https://api.openai.com/v1

以下所有示例均假设你已设置好

OPENAI_API_KEY

。

基础URL：

https://api.openai.com/v1

1. Basic Chat Completion

1. 基础对话补全

Send a simple chat message:

Write to

/tmp/openai_request.json

json

{
  "model": "gpt-4o-mini",
  "messages": [{"role": "user", "content": "Hello, who are you?"}]
}

Then run:

bash

bash -c 'curl -s "https://api.openai.com/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json' | jq '.choices[0].message.content'

Available models:

```
gpt-4o
```
: Latest flagship model (128K context)
```
gpt-4o-mini
```
: Fast and affordable (128K context)
```
gpt-4-turbo
```
: Previous generation (128K context)
```
gpt-3.5-turbo
```
: Legacy model (16K context)
```
o1
```
: Reasoning model for complex tasks
```
o1-mini
```
: Smaller reasoning model

发送简单的对话消息：

写入到

/tmp/openai_request.json

：

json

{
  "model": "gpt-4o-mini",
  "messages": [{"role": "user", "content": "Hello, who are you?"}]
}

然后运行：

bash

bash -c 'curl -s "https://api.openai.com/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json' | jq '.choices[0].message.content'

可用模型：

```
gpt-4o
```
: 最新旗舰模型（128K上下文）
```
gpt-4o-mini
```
: 快速且经济实惠（128K上下文）
```
gpt-4-turbo
```
: 上一代模型（128K上下文）
```
gpt-3.5-turbo
```
: 旧版模型（16K上下文）
```
o1
```
: 用于复杂任务的推理模型
```
o1-mini
```
: 轻量化推理模型

2. Chat with System Prompt

2. 带系统提示的对话

Use a system message to set behavior:

Write to

/tmp/openai_request.json

json

{
  "model": "gpt-4o-mini",
  "messages": [
    {"role": "system", "content": "You are a helpful assistant that responds in JSON format."},
    {"role": "user", "content": "List 3 programming languages with their main use cases."}
  ]
}

Then run:

bash

bash -c 'curl -s "https://api.openai.com/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json' | jq '.choices[0].message.content'

使用系统消息设置助手行为：

写入到

/tmp/openai_request.json

：

json

{
  "model": "gpt-4o-mini",
  "messages": [
    {"role": "system", "content": "You are a helpful assistant that responds in JSON format."},
    {"role": "user", "content": "List 3 programming languages with their main use cases."}
  ]
}

然后运行：

bash

bash -c 'curl -s "https://api.openai.com/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json' | jq '.choices[0].message.content'

3. Streaming Response

3. 流式响应

Get real-time token-by-token output:

Write to

/tmp/openai_request.json

json

{
  "model": "gpt-4o-mini",
  "messages": [{"role": "user", "content": "Write a haiku about programming."}],
  "stream": true
}

Then run:

bash

curl -s "https://api.openai.com/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json

Streaming returns Server-Sent Events (SSE) with delta chunks.

获取实时的逐token输出：

写入到

/tmp/openai_request.json

：

json

{
  "model": "gpt-4o-mini",
  "messages": [{"role": "user", "content": "Write a haiku about programming."}],
  "stream": true
}

然后运行：

bash

curl -s "https://api.openai.com/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json

流式响应返回带delta块的Server-Sent Events（SSE）。

4. JSON Mode

4. JSON模式

Force the model to return valid JSON:

Write to

/tmp/openai_request.json

json

{
  "model": "gpt-4o-mini",
  "messages": [
    {"role": "system", "content": "Return JSON only."},
    {"role": "user", "content": "Give me info about Paris: name, country, population."}
  ],
  "response_format": {"type": "json_object"}
}

Then run:

bash

bash -c 'curl -s "https://api.openai.com/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json' | jq '.choices[0].message.content'

强制模型返回有效的JSON：

写入到

/tmp/openai_request.json

：

json

{
  "model": "gpt-4o-mini",
  "messages": [
    {"role": "system", "content": "Return JSON only."},
    {"role": "user", "content": "Give me info about Paris: name, country, population."}
  ],
  "response_format": {"type": "json_object"}
}

然后运行：

bash

bash -c 'curl -s "https://api.openai.com/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json' | jq '.choices[0].message.content'

5. Vision (Image Analysis)

5. 视觉功能（图像分析）

Analyze an image with GPT-4o:

Write to

/tmp/openai_request.json

json

{
  "model": "gpt-4o-mini",
  "messages": [
    {
      "role": "user",
      "content": [
        {"type": "text", "text": "What is in this image?"},
        {"type": "image_url", "image_url": {"url": "https://upload.wikimedia.org/wikipedia/commons/thumb/3/3a/Cat03.jpg/1200px-Cat03.jpg"}}
      ]
    }
  ],
  "max_tokens": 300
}

Then run:

bash

bash -c 'curl -s "https://api.openai.com/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json' | jq '.choices[0].message.content'

使用GPT-4o分析图像：

写入到

/tmp/openai_request.json

：

json

{
  "model": "gpt-4o-mini",
  "messages": [
    {
      "role": "user",
      "content": [
        {"type": "text", "text": "What is in this image?"},
        {"type": "image_url", "image_url": {"url": "https://upload.wikimedia.org/wikipedia/commons/thumb/3/3a/Cat03.jpg/1200px-Cat03.jpg"}}
      ]
    }
  ],
  "max_tokens": 300
}

然后运行：

bash

bash -c 'curl -s "https://api.openai.com/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json' | jq '.choices[0].message.content'

6. Function Calling (Tools)

6. 函数调用（工具）

Define functions the model can call:

Write to

/tmp/openai_request.json

json

{
  "model": "gpt-4o-mini",
  "messages": [{"role": "user", "content": "What is the weather in Tokyo?"}],
  "tools": [
    {
      "type": "function",
      "function": {
        "name": "get_weather",
        "description": "Get current weather for a location",
        "parameters": {
          "type": "object",
          "properties": {
            "location": {"type": "string", "description": "City name"}
          },
          "required": ["location"]
        }
      }
    }
  ]
}

Then run:

bash

bash -c 'curl -s "https://api.openai.com/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json' | jq '.choices[0].message.tool_calls'

定义模型可调用的函数：

写入到

/tmp/openai_request.json

：

json

{
  "model": "gpt-4o-mini",
  "messages": [{"role": "user", "content": "What is the weather in Tokyo?"}],
  "tools": [
    {
      "type": "function",
      "function": {
        "name": "get_weather",
        "description": "Get current weather for a location",
        "parameters": {
          "type": "object",
          "properties": {
            "location": {"type": "string", "description": "City name"}
          },
          "required": ["location"]
        }
      }
    }
  ]
}

然后运行：

bash

bash -c 'curl -s "https://api.openai.com/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json' | jq '.choices[0].message.tool_calls'

7. Generate Embeddings

7. 生成文本嵌入

Create vector embeddings for text:

Write to

/tmp/openai_request.json

json

{
  "model": "text-embedding-3-small",
  "input": "The quick brown fox jumps over the lazy dog."
}

Then run:

bash

bash -c 'curl -s "https://api.openai.com/v1/embeddings" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json' | jq '.data[0].embedding[:5]'

This extracts the first 5 dimensions of the embedding vector.

Embedding models:

```
text-embedding-3-small
```
: 1536 dimensions, fastest
```
text-embedding-3-large
```
: 3072 dimensions, most capable

为文本创建向量嵌入：

写入到

/tmp/openai_request.json

：

json

{
  "model": "text-embedding-3-small",
  "input": "The quick brown fox jumps over the lazy dog."
}

然后运行：

bash

bash -c 'curl -s "https://api.openai.com/v1/embeddings" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json' | jq '.data[0].embedding[:5]'

此命令提取嵌入向量的前5个维度。

嵌入模型：

```
text-embedding-3-small
```
: 1536维度，速度最快
```
text-embedding-3-large
```
: 3072维度，功能最强大

8. Generate Image (DALL-E 3)

8. 生成图像（DALL-E 3）

Create an image from text:

Write to

/tmp/openai_request.json

json

{
  "model": "dall-e-3",
  "prompt": "A white cat sitting on a windowsill, digital art",
  "n": 1,
  "size": "1024x1024"
}

Then run:

bash

bash -c 'curl -s "https://api.openai.com/v1/images/generations" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json' | jq '.data[0].url'

Parameters:

```
size
```
:
```
1024x1024
```
,
```
1792x1024
```
, or
```
1024x1792
```
```
quality
```
:
```
standard
```
or
```
hd
```
```
style
```
:
```
vivid
```
or
```
natural
```

根据文本描述创建图像：

写入到

/tmp/openai_request.json

：

json

{
  "model": "dall-e-3",
  "prompt": "A white cat sitting on a windowsill, digital art",
  "n": 1,
  "size": "1024x1024"
}

然后运行：

bash

bash -c 'curl -s "https://api.openai.com/v1/images/generations" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json' | jq '.data[0].url'

参数说明：

```
size
```
:
```
1024x1024
```
、
```
1792x1024
```
或
```
1024x1792
```
```
quality
```
:
```
standard
```
（标准）或
```
hd
```
（高清）
```
style
```
:
```
vivid
```
（生动）或
```
natural
```
（自然）

9. Audio Transcription (Whisper)

9. 音频转写（Whisper）

Transcribe audio to text:

bash

bash -c 'curl -s "https://api.openai.com/v1/audio/transcriptions" -H "Authorization: Bearer ${OPENAI_API_KEY}" -F "file=@audio.mp3" -F "model=whisper-1"' | jq '.text'

Supports: mp3, mp4, mpeg, mpga, m4a, wav, webm (max 25MB).

将音频转换为文本：

bash

bash -c 'curl -s "https://api.openai.com/v1/audio/transcriptions" -H "Authorization: Bearer ${OPENAI_API_KEY}" -F "file=@audio.mp3" -F "model=whisper-1"' | jq '.text'

支持的格式：mp3、mp4、mpeg、mpga、m4a、wav、webm（最大25MB）。

10. Text-to-Speech

10. 文本转语音

Generate audio from text:

Write to

/tmp/openai_request.json

json

{
  "model": "tts-1",
  "input": "Hello! This is a test of OpenAI text to speech.",
  "voice": "alloy"
}

Then run:

bash

curl -s "https://api.openai.com/v1/audio/speech" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json --output speech.mp3

Voices:

alloy

echo

fable

onyx

nova

shimmer

Models:

tts-1

(fast),

tts-1-hd

(high quality)

根据文本生成音频：

写入到

/tmp/openai_request.json

：

json

{
  "model": "tts-1",
  "input": "Hello! This is a test of OpenAI text to speech.",
  "voice": "alloy"
}

然后运行：

bash

curl -s "https://api.openai.com/v1/audio/speech" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json --output speech.mp3

可用音色：

alloy

、

echo

、

fable

、

onyx

、

nova

、

shimmer

可用模型：

tts-1

（快速）、

tts-1-hd

（高质量）

11. List Available Models

11. 列出可用模型

Get all available models:

bash

bash -c 'curl -s "https://api.openai.com/v1/models" -H "Authorization: Bearer ${OPENAI_API_KEY}"' | jq -r '.data[].id' | sort | head -20

获取所有可用的模型：

bash

bash -c 'curl -s "https://api.openai.com/v1/models" -H "Authorization: Bearer ${OPENAI_API_KEY}"' | jq -r '.data[].id' | sort | head -20

12. Check Token Usage

12. 查看Token使用情况

Extract usage from response:

Write to

/tmp/openai_request.json

json

{
  "model": "gpt-4o-mini",
  "messages": [{"role": "user", "content": "Hi!"}]
}

Then run:

bash

bash -c 'curl -s "https://api.openai.com/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json' | jq '.usage'

This returns token counts for both input and output.

Response includes:

```
prompt_tokens
```
: Input token count
```
completion_tokens
```
: Output token count
```
total_tokens
```
: Sum of both

从响应中提取使用数据：

写入到

/tmp/openai_request.json

：

json

{
  "model": "gpt-4o-mini",
  "messages": [{"role": "user", "content": "Hi!"}]
}

然后运行：

bash

bash -c 'curl -s "https://api.openai.com/v1/chat/completions" -H "Content-Type: application/json" -H "Authorization: Bearer ${OPENAI_API_KEY}" -d @/tmp/openai_request.json' | jq '.usage'

此命令返回输入和输出的Token数量。

响应包含：

```
prompt_tokens
```
: 输入Token数量
```
completion_tokens
```
: 输出Token数量
```
total_tokens
```
: 总Token数量

Guidelines

使用指南

Choose the right model: Use
```
gpt-4o-mini
```
for most tasks,
```
gpt-4o
```
for complex reasoning,
```
o1
```
for advanced math/coding
Set max_tokens: Prevent runaway generation and control costs
Use streaming for long responses: Better UX for real-time applications
JSON mode requires system prompt: Include JSON instructions when using
```
response_format
```
Vision requires gpt-4o models: Only
```
gpt-4o
```
and
```
gpt-4o-mini
```
support image input
Batch similar requests: Use embeddings API batch input for efficiency
Monitor usage: Check dashboard regularly to avoid unexpected charges

选择合适的模型：大多数任务使用
```
gpt-4o-mini
```
，复杂推理使用
```
gpt-4o
```
，高级数学/编码任务使用
```
o1
```
设置max_tokens：防止生成内容过长并控制成本
长响应使用流式输出：为实时应用提供更好的用户体验
JSON模式需要系统提示：使用
```
response_format
```
时需包含JSON格式说明
视觉功能仅支持GPT-4o系列模型：只有
```
gpt-4o
```
和
```
gpt-4o-mini
```
支持图像输入
批量处理相似请求：使用嵌入API的批量输入提高效率
监控使用情况：定期查看仪表盘以避免意外收费